1. Field
The present disclosure generally relates to forming relationships between image features.
2. Background
Various methods exist for extracting features from images. Examples of feature detection algorithms include scale-invariant feature transform (“SIFT”), difference of Gaussians, maximally stable extremal regions histogram of oriented gradients, gradient location and orientation histogram, smallest univalue segment assimilating nucleus, and speeded-up robust features. Also, multiple features are sometimes used to recognize objects in images. Additionally, an image may be associated with one or more labels, for example tags that are applied by a user.